qiyd81 / botlab

Project for ROB 550. Implementing SLAM using LIDAR, wheel encoders and IMU on MBot.

Slam and Path Planning implementation on MBot

A project for ROB 550: Robotics Systems Lab course taught in University of Michigan, Ann Arbor. Due to the COVID-19 pandemic this project was migrated to an online project. An overview of this project:

• Acting

  • Planar kinematics of a differential-drive ground robot
  • Motion models with uncertainty

• Perception

  • Quadrature Encoders
  • MEMS Inertial Measurement Unit
  • 2D LIDAR Rangefinders

• Reasoning

  • Monte Carlo Localization
  • Simultaneous Localization and Mapping
  • A* search
  • Path planning

Running the code

log_mbot_sensors.sh

• a script to log the sensor data needed for SLAM so you can easily create your own log files for testing

setenv.sh

• a script for setting environment variables needed for running Vx applications
• run this script before running botgui in a terminal on your laptop
• run via . setenv.sh -- note the space

1. Mapping

   • lcm-logplayer-gui <log_file.log> and uncheck all SLAM channels but SLAM_POSE
   • ./botgui
   • ./slam --mapping-only

2. Action Model

   • lcm-logplayer-gui <log_file.log> and uncheck all SLAM channels
   • ./botgui
   • ./slam --action-only

3. Sensor Model

   • lcm-logplayer-gui <log_file.log> and uncheck all SLAM channels
   • ./botgui
   • ./slam --localization-only <map_file.map>

4. Full SLAM

   • lcm-logplayer-gui <log_file.log> and uncheck all SLAM channels
   • ./botgui
   • ./slam

5. Obstacle Distance Grid

   • lcm-logplayer-gui <log_file.log> and uncheck all SLAM channels
   • ./botgui
   • View the generated obstacle distance grid by ticking “Show Obstacle Distances” in botgui

6. A* Algorithm

   • cd <path_to_dir>/src/sim
   • python3 sim.py <map_file.map>
   • ./timesync
   • ./botgui
   • ./motion_controller
   • ./slam --localization-only <map_file.map>
   • Right-click on the botgui to drive the robot to that location

7. Frontier exploration

   • cd <path_to_dir>/src/sim
   • python3 sim.py <map_file.map>
   • ./timesync
   • ./botgui
   • ./motion_controller
   • ./slam
   • ./exploration

Check the final report for detailed explanation and results.

Directories

bin/

• where all built binaries are located

data/

• where data needed to run parts of the assignment are located
• log files and target files for SLAM and exploration are here

lcmtypes/

• where the .lcm type definitions are located
• the generated types are stored in src/lcmtypes

lib/

• where static libraries are saved during the build process

src/

• where all source code for botlab is located
• subdirectories have a further description of their contents